Non-small cell lung cancer (NSCLC) is a significant health concern. The prognostic value of oxidative stress (OS)-related genes in NSCLC remains unclear. The study aimed to explore the prognostic significance of OS-genes in NSCLC using extensive datasets from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO).

This study successfully developed a robust model for predicting patient prognosis in NSCLC, highlighting the critical prognostic value of OS-genes. These findings hold significant potential to refine treatment strategies, and enhance survival outcomes for NSCLC patients. By enabling a personalized therapeutic approach tailored to individual risk scores, this model may facilitate more precise decisions concerning immunotherapy and chemotherapy, thereby optimizing patient management and treatment efficacy.

The research used the expression data and clinical information of NSCLC patients to develop a risk-score model. A total of 74 OS-related differentially expressed genes (DEGs) were identified by comparing NSCLC and control samples. Univariate Cox and least absolute shrinkage and selection operator (LASSO) regression analyses were employed to identify the prognostic biomarkers. A risk-score model was constructed and validated with receiver operating characteristic (ROC) curves in TCGA and GSE72094 datasets. The model's accuracy was further verified by univariate and multivariate Cox regression.

The identified biomarkers, including lactate dehydrogenase A (LDHA), protein tyrosine phosphatase receptor type N (PTPRN), and transient receptor potential cation channel subfamily A (TRPA1) demonstrated prognostic significance in NSCLC. The risk-score model showed good predictive accuracy, with 1-year area under the curves (AUC) of 0.661, 3-year AUC of 0.648, and 5-year AUC of 0.634 in the TCGA dataset, and 1-year AUC of 0.643, 3-year AUC of 0.648, and 5-year AUC of 0.662 in the GSE72094 dataset. A nomogram integrating risk score and tumor node metastasis (TNM) stage was developed. The signature effectively distinguished between patient responses to immunotherapy. High-risk groups were characterized by an immunosuppressive microenvironment and an increased tumor mutational burden (TMB), marked by a higher incidence of mutations in genes such as TP53, DCP1B, ELN, and MAGI2. Organoid drug sensitivity testing revealed that NSCLC patients with a low-risk score responded better to chemotherapy. Conclusions: This study successfully developed a robust model for predicting patient prognosis in NSCLC, highlighting the critical prognostic value of OS-genes. These findings hold significant potential to refine treatment strategies, and enhance survival outcomes for NSCLC patients. By enabling a personalized therapeutic approach tailored to individual risk scores, this model may facilitate more precise decisions concerning immunotherapy and chemotherapy, thereby optimizing patient management and treatment efficacy.